This resonates- my engineering workflow has started shifting from highly focused, long periods of building out a feature to one that has much more context switching, review, and testing.
In case this sparked anybody's interest, I went through the tweet's author's course on RL [1] recently and it was extremely insightful. I heartily recommend it (slides, lectures, assignments are all public)
Hey! I happen to be building something very similar in my free time- started off as a hackathon project. Feel free to reach out! Only spent around 10 hours on the current deployed iteration but you can check the idea out here- https://www.withtal.com/
I ran into the same issue out of nowhere with my Spotify app a few months ago- after hitting around 300 users. Apparently, Spotify has a limit [0] on their API quota for auth (around 25 users or so in theory), but don't feature the limit very prominently in their documentation.
Have you requested and obtained a quota extension? I gave up on my app altogether after a couple of weeks and seeing in their documentation "[this] review process can take up to six weeks."
This seems like a great practical framework- I've noticed that establishing habits like the ones you mention are the only reliable first step to long projects in my life.
Very curious about what you would do on a daily bases that would qualify as "interesting"? Any particularly rewarding "interesting" activities? I'd love to inject some routine variety in my life.
Agree on relatively strict self-selection being a necessary component for any thriving community. Also, most (dare I say all?) of my friendships have been formed not because I stumbled across an interesting person I decided to bond with, but because some external circumstance (living situation, hobbies) made it so that we were in close contact over an extended period of time.
I've thought a bit about this issue and technology's potential role in fixing it. A generalized app that promises community isn't going to reach mass adoption outside of an extremely viral built-in mechanism. Mostly because the required user density is inevitably so high (also why most new dating apps fail). If you (or any other entrepreneur) is focused on tackling this issue, I'd do it incrementally, by fixing a clear problem in one very specific segment of your potential audience's life, and expanding from there.
Look at an website like https://www.lu.ma which is in some sense accomplishing a similar mission.
What would you have them study instead? If you're optimizing for what AI is least likely to be able to accomplish, some forms of blue collar labor would probably rank highest on the list.
Which restrictions specifically? What's your use case? I'm curious because I've completely shelved some projects when I realized I wouldn't be able to prompt GPT to respond in a natural way to specific classes of prompts- ie. some classes of writing.
> Off topic: blog posts, sign-up pages, newsletters, lists, and other reading material. Those can't be tried out, so can't be Show HNs. Make a regular submission instead.
Fascinating that the human benchmark is 63%- I wonder what the benchmark would look like were it to have been established, say, 30 years ago, before the prevalence of LLMs; I'd wager it would be very close to 100%. Speaks to the moving goalposts.
> You know, years ago I wrote a little program to look at this, like how quickly our best founders — the founders that run billion-plus companies — answer my emails versus our bad founders. I don’t remember the exact data, but it was mind-blowingly different. It was a difference of minutes versus days on average response times